Skip to content

News

Security incident

Researchers bypass ChatGPT Lockdown Mode; OpenAI patched the flaw

Check Point Research found a way around ChatGPT's Lockdown Mode that could let chat data leave the app without user interaction. OpenAI fixed it after responsible disclosure.

Check Point Research (CPR) has found a way to bypass ChatGPT's Lockdown Mode, the optional setting meant to stop prompt injection from moving sensitive data out of the chat. According to CPR, the bypass could enable silent chat data theft, phishing and malware delivery, including zero-click data exfiltration in the ChatGPT desktop application.

Lockdown Mode restricts several outbound channels, such as live web access and internet-hosted images, and disables features like Deep Research, Agent Mode, live connectors and file downloads. CPR found that another feature in the application offered an alternative path around these limits.

OpenAI fixed the vulnerability after responsible disclosure. CPR confirmed that the technique described in its research no longer works. For users of AI chat services, the case shows that protections built into the model or its settings depend on the wider application around them.

Reported on 6 November 2026 by TechRadar, based on the Check Point Research blog post.

Sources